Re: iSCSI - started countdown to 12 (11-95)

["Michael J. S. Smith \(RoadRunner\)" <smithm007@hawaii.rr.com>]

Julo: This is a really silly one, but I accidentally found a
typographical error while reading 11-95, section 7.3.1:

An iSCSI compliant initiator or target MUST provide data integrity and
authentication by implementing IPsec [RFC2401] with ESP in tunnel mode
[RFC2406] and MAY provide data provide data integrity and authen-
tication by implementing IPsec with ESP in transport mode. The IPsec
implementation MUST fulfill the following iSCSI specific require- ments:

"provide data provide data" -> "provide data"

I wouldn't normally waste bandwidth on this, but there's an interesting
discussion going on over in the SNIA snia-ips forum on iSCSI IPsec APIs
(which does not belong here) and while digging around 11-95 I noticed
the above. Anyway, during the discussion, while Bernard was pointing out
some of the wording in the ips-security draft, it occured to me that I
needed to ask David (and Julo and Bernard, I guess) the following
procedural question:

David, Julo, and Bernard: do we, and if so, how do we ensure that the
iSCSI Internet draft and ips-security drafts are true to each other?

The ambiguity in using the phrase "true to each other" was deliberate,
how do these things normally work, David? Does the ips-security draft
lapse or does it become informational or do we just keep leapfrogging or
what (I did read RFC1796 and RFC2026)? As a small concrete example, how
do we cover the things going on in AES counter mode in saag and the OCB
and CCM work in 802.11 TGi? Do we expect to update the iscsi draft, the
ips-security draft, or both, or neither?

Aloha
Mike Smith
CTO, iReady
(I'm currently in Hawaii, ignore my email address, reply to
msmith@iready.com)
Encoding: Western European (ISO), apologies for those expecting plain
ASCII.

----- Original Message -----
From: "Julian Satran" <Julian_Satran@il.ibm.com>
To: <ips@ece.cmu.edu>
Sent: Saturday, April 13, 2002 2:58 AM
Subject: iSCSI - started countdown to 12 (11-95)


> Dear colleagues,
>
> I've put on my site (http://www.haifa.il.ibm.com/satran/ips) a
"working" version of the draft labeled 11-95
> Only the pdf version (with change bars vs. 11) is available.
> It contains all the agreed changes + a MUST requirement for the
initiator
> to fully deliver data on R2T and
> unsolicited data-out (thanks Ralph Weber for the convincing
arguments).
> security - tunnel MUST transport MAY  and authenticate when encrypting
> the normative naming text
> the last version of the clearing effects appendix
> removed CRN
> an implementer note about TPGT usage and discovery
> unsolicited data - if not immediate or none must be of FirstBufSize or
> Total Data (I am not sure we should allow immediate but I see no real
> harm).
> references are split -
> minor corrections in text and AE data has now the same format as
response
>
> Julo
>
>
>
>
>
>
>
>
>
>
>
>
>
>